Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752400AbcLER6a (ORCPT ); Mon, 5 Dec 2016 12:58:30 -0500 Received: from mailapp01.imgtec.com ([195.59.15.196]:49373 "EHLO mailapp01.imgtec.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752330AbcLER63 (ORCPT ); Mon, 5 Dec 2016 12:58:29 -0500 Date: Mon, 5 Dec 2016 17:41:49 +0000 From: "Maciej W. Rozycki" To: Paul Burton CC: Matt Redfearn , Ralf Baechle , , "Jason A . Donenfeld" , Thomas Gleixner , , James Hogan Subject: Re: [PATCH 3/5] MIPS: Only change $28 to thread_info if coming from user mode In-Reply-To: <40031221.jIbnRxB01Q@np-p-burton> Message-ID: References: <1480685957-18809-1-git-send-email-matt.redfearn@imgtec.com> <1480685957-18809-4-git-send-email-matt.redfearn@imgtec.com> <40031221.jIbnRxB01Q@np-p-burton> User-Agent: Alpine 2.00 (DEB 1167 2008-08-23) MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" X-Originating-IP: [10.20.78.176] Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 765 Lines: 16 On Mon, 5 Dec 2016, Paul Burton wrote: > Agreed we ought to use .set reorder (or rather, not use .set noreorder) > wherever possible but FYI one thing I've only noticed recently is that we > don't actually get any reordering anyway, presumably because we don't provide > any -O flags when building asm source. I haven't yet done the legwork or > figuring out whether we've ever had optimisation & if so what changed... Reordering or `-O2' is the default for GAS, you actually have to pass a different `-O*' option explicitly to GAS to disable it. NB it's *not* the same as passing one of the `-O*' options to GCC, you'd need e.g. `-Wa,-O'. So perhaps one of our Makefiles or a particular configuration of the GCC driver does the wrong thing? Maciej